JP2005131940A - Manufacturing method for matrix for microlens array - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a manufacturing method for a matrix for a microlens array, which enables the manufacture of the matrix for the microlens array, composed of a unit cell with curvature controlled with an extremely high degree of accuracy, on a conventionally unrealizable large-area substrate by solving the problem of roughening of a substrate surface, as a defect in laser ablation. <P>SOLUTION: In this manufacturing method for the matrix for forming an optical member provided with the many microlens arrays, a photomask 2, where a mask pattern is formed, is arranged above the substrate 1; a laser beam 3 is applied via the photomask 2, so that a desired three-dimensional pattern depending on the mask pattern can be formed on the substrate 1; in this case, the spot size of the laser beam 3 is increased so that the laser beam can pass through the photomask 2; and subsequently, the spot size of the laser beam 3 is reduced for the application of the laser beam 1 to the surface of the substrate 1, so that the surface of the substrate can be partially evaporated. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a method for manufacturing a matrix for a microlens array, and more particularly to a method for manufacturing a large area matrix used to manufacture a microlens array sheet that forms part of a screen for a rear projection display. is there.

  In general, a rear projection display device mainly includes three types of components: an optical engine that is a component that generates an image, an optical system that projects image light emitted therefrom onto a projection screen, and a projection screen that receives the image light. Consists of.

  Among these, the projection screen closest to the observer is composed of two types of optical members, which are generally called Fresnel lenses and lenticular plates from the optical engine side. The function of the Fresnel lens is collimation of the entire image light, and the function of the lenticular plate is adjustment of the diffusion angle of the pixel light emitted from each pixel.

  When the image light emitted from the optical engine reaches the projection screen, it is magnified by the optical system so that it has a cross section that matches the effective display area of the projection screen. Although all the image light emitted from each pixel is diffused in the same direction. Since the direction in which the intensity of the emitted image light is maximum differs for each pixel, when using a projection screen that does not have a pixel light diffusion angle adjustment function, it is possible for an observer viewing from any direction. The same display quality cannot be shown over the entire display screen. This is because the viewing angle-luminance distribution characteristics are different for each pixel.

  By the way, for data display, for example, a computer monitor, there is a design philosophy that it is necessary to show the same display quality over the entire screen to any observer who is viewing from an effective viewing angle range. For this purpose, it is necessary that the intensity of the individual pixel light emitted from all the pixels constituting the screen is diffused so as to be the same in any emission angle direction within the effective viewing angle. When such a diffusing surface is used, an observer who is located at the boundary between the inside and outside of the effective viewing angle range will see a sudden change in display quality with only a slight movement of the viewpoint. In other words, it is considered that there is no problem because there is only one observer.

  There is also a design philosophy that it is preferable that the display quality gradually changes depending on the position of the observer within the effective viewing angle range, which is considered suitable for use by a large number of people, that is, for a television. . In most cases, the diffusion pattern is such that the brightness is the maximum when viewed from the front, that is, the viewing angle (the angle formed by the perpendicular to the screen and the viewer's line of sight) = 0 °, and gradually decreases as the viewing angle tilts. Design is made. Many of the projection displays that can be seen on the market today are designed with these characteristics in mind. Also in this case, it is necessary to make the diffusion pattern of the emitted light from all the pixels the same.

  In order to make the diffusion pattern of each pixel light emitted from all the pixels on the projection screen the same, it reaches the projection screen surface from the optical engine by the magnifying optical system at various incident angles and has different luminous intensity. Correction using an optical system must be performed for each pixel light having an angular distribution of λ.

  That is, the above object can be achieved by providing a microlens for correcting the optical axis (the direction indicating the maximum luminous intensity) and the diffusion angle at each position corresponding to each pixel on the projection screen surface.

  However, since it is necessary to perform a very precise positioning operation to accurately place the microlens at positions corresponding to a large number of pixels on the projection screen surface, the equipment is expensive and the productivity is low. Therefore, the manufacturing cost becomes excessive.

  In addition, since such a microlens array is designed for a combination of a specific optical system and an optical engine, it has no versatility, so that the manufacturing cost further increases.

  Therefore, in order to avoid the positioning operation of the pixel on the optical engine side and the pixel on the projection screen surface, the microlens arranged on the projection screen surface is generally divided into a plurality of parts.

  At this time, if the pitch of the microlenses arranged on the projection screen is set to 1/5 or less of the pixel pitch, it is considered that no visible decrease in resolution occurs even if a positional deviation occurs to some extent.

  On the other hand, it is extremely complicated and difficult to design and manufacture microlenses that correct the optical axis and diffusion angle of the light emitted from each pixel on the projection screen surface. Currently, a method is employed in which the entire incident light is converted into parallel light, and the optical axis correction is not performed in the microlens array.

  That is, the Fresnel lens is first arranged on the optical engine side of the projection screen to perform collimation, and the diffusion angle of the parallel light emitted from the Fresnel lens is corrected by the microlens array. Since the image light incident on the microlens array is parallel light, it is the most important function of the microlens array to widen the outgoing light with an angular distribution, that is, to correct the diffusion angle. The angular distribution is determined by the curved surface shape of the unit cell constituting the microlens array.

  As described above, all of the projection display screens have this configuration.

  However, currently used microlens arrays still have serious drawbacks. A microlens array used for a projection display screen is a cylindrical lens array generally called a lenticular plate, and has a diffusion angle correction function only in one axial direction. The main application of the current rear projection display is a TV, which is often viewed by a large number of people, so the direction that requires diffusion angle correction is mainly the horizontal direction, but it is not without the need for correction in the vertical direction. Since it is not required as much as in the lateral direction, the longitudinal direction of the cylindrical lens unit cell is used vertically.

  The emission angle in the vertical direction can only be slightly expanded by adding fine particles with different refractive indexes to the members constituting the cylindrical lens array to give scattering characteristics. is the current situation.

  In this state of the art, there are three problems shown in the following a, b and c.

aVertical viewing angle is narrow. Despite attempts to increase the vertical viewing angle by providing scattering properties, the viewing angle can hardly be increased by scattering. In many cases, which are used to display data in public places, the display is often placed higher than the observer's eyes, and the narrow vertical viewing angle is a serious problem.

(b) One of the serious side effects caused by scattering by fine particles added for increasing the vertical viewing angle is a reduction in resolution. This is not a serious problem for large-area TVs, which are the main applications of current projection displays, but it will be a fatal problem if the resolution increases in the future.

cAnother side effect of scattering by fine particles is speckle (fine luminescent spots).
Occurs. This phenomenon is caused by scintillation due to the presence of many scatterers that are smaller than the coherent length of the optical system.

  In order to avoid these problems, two cylindrical lenses are crossed and used. However, in this case as well, there are two problems indicated by a and b below.

aSince two cylindrical lens array sheets are used, the manufacturing cost of this part is simply more than doubled, resulting in poor economic efficiency.

b The interface between the air and the member becomes four places, the reflection loss increases, and the display quality deteriorates due to unexpected imaging caused by the reflected light becoming stray light. The use of anti-reflection coatings increases manufacturing costs.

  Therefore, in order to obtain good display quality in the rear projection display, it is essential that no scatterers are added, and the number of microlens array sheets to be used must be limited to only one.

  For this purpose, a microlens array having condensing characteristics in two axial directions is necessary. Such a macro lens array must have a structure in which minute spherical or aspherical lenses, generally called fly-eye lenses, are arranged vertically and horizontally instead of cylindrical lenses, and the microlens unit cell that constitutes them is The curved surface shape must be controlled so that the designed emission light diffusion characteristics can be obtained. Desirable curved surface shapes are often paraboloids, hyperboloids expressed by specific parameters, or toric surfaces with different vertical and horizontal curvatures.

  However, there has been no projection display screen using a microlens array sheet composed of such fly-eye lenses. This is because it was extremely difficult to produce a mold for molding.

  In the case of a cylindrical lens array, processing in one axis direction is sufficient for manufacturing a mold, so it is possible to manufacture by cutting a groove having a desired cross-sectional shape on the surface of a metal base material using a lathe. Is possible. Once the master mold has been successfully manufactured, the plastic member may be molded by pressing, casting, or extrusion.

  However, in the case of a fly-eye lens, since the structure exists in two or more directions, it is impossible to process with a lathe. Although it is not without machining means, for example, if the unit cell is cut one by one using an end mill, the center point of processing is at the center of the unit cell in addition to the cost increase due to the time required. There is a fatal defect that it remains in an abnormal shape.

  Thus, a method for manufacturing a fly-eye lens without machining has been developed so far. They are shown in the following 1-4.

1 Array of rigid microspheres
Microspheres of the same size as the unit cell are arranged on a plane and fixed. Capable of handling large areas
However, when the unit cell size becomes small, it becomes difficult to cope with it. Array operation
Because of this problem, it is impossible to select a unit cell shape other than spherical, that is,
In other words, it is impossible to align the direction of the shape when something other than a sphere is arranged
The desired pixel light diffusion characteristics cannot be obtained.
.

2 Curved surface formation by liquid surface tension (1) Reflow method
Pattern photoresist into a cylindrical shape or polymer film by laser processing
A cylindrical array is formed and heated to flow to obtain a curved surface.
Large area is possible, but unit cell shape is determined by resist flow
Therefore, there is no degree of freedom in design, and a desirable pixel light diffusion characteristic cannot be obtained. Ma
When a polymer composing a cylinder softens and flows, it flows from an adjacent cylinder.
When it comes into contact with polymer, it is connected smoothly due to surface tension, and the shape is
It will be distorted.

(2) Inkjet method
The polymer solution is ejected by an ink jet printer and adhered onto the substrate,
A curved surface is obtained by flow.
Advantages and disadvantages are the same as the reflow method in the previous section.

3 Partial etching of rigid plane
Attacked by chemicals that dissolve this transparent rigid plate material on the surface of transparent rigid plates such as glass
A thin film made of a material that cannot be
The transparent rigid flat plate surface is etched through the hole.
It is possible to deal with large areas, but etching starting from holes progress isotropically.
The unit cell shape is spherical, and there is no degree of freedom in design.
Can't get.

4 Three-dimensional lithography (1) X-ray lithography
Usually, X-ray lithography is LIGA (LIthography-Galvan
Ischversilbern-Abduck)
Used to create 3D patterns that are very deep
Application to turn formation has begun recently.
A mask made from a cross-sectional shape cut along the symmetry axis of a fly-eye lens.
X-rays while sliding in a direction perpendicular to the axis and parallel to the resist-coated substrate
Irradiation is performed for each axis of symmetry, and the resist is developed.
Highly controlled unit cell curved surface shape can be created, but it can handle large areas.
It ’s difficult. In other words, since the synchrotron is used for the radiation source, the equipment cost
In addition to being extremely large, the substrate is precisely positioned with respect to the radiation source.
Currently, there is no equipment that can perform ring exposure.

(2) Electron beam lithography a Direct method
Irradiation is performed while scanning the electron beam onto the surface of the substrate coated with the electron beam resist.
Adjust the dose according to the pattern in the depth direction of the lens array and absorb the resist.
After developing the reactive ion energy, a bias voltage was applied perpendicularly to the resist coating surface.
Etch the substrate by etching. Double determined by selectivity with electron beam resist
The three-dimensional pattern of the resist whose depth dimension is amplified by
It is copied.
Highly controlled unit cell curved surface shape can be created, but it can handle large areas.
It ’s difficult. That is, in general, the absorbed electron dose is modulated by the electron beam being scanned.
Realized by performing ON-OFF in multiple scans, but the board area is
The larger the time, the more time it takes to scan, and the completion within a realistic time
I can't hope. Also, the same scale as the projection display screen
Currently, there is no facility capable of scanning an electron beam over a certain area.

Indirect method
While scanning an electron beam on the surface of a glass substrate that is colored by absorbing the electron beam
Irradiate. Modulate and absorb the dose according to the depth pattern of the lens array
Based on the gray scale pattern in which the depth pattern is converted to shades of color
Draw on a board. Use this as a photomask to create a photoresist with light in the wavelength range
The resist-coated substrate is exposed and developed to create a resist three-dimensional pattern.
It is formed on the plate surface.
Highly controlled unit cell curved surface shape can be created, but it can handle large areas.
It ’s difficult. The reason is the same as in the direct method.
However, the steps used to form a fine pattern on a large area LCD substrate.
The hopper is now almost the same size as the screen of a projection display
The ones that can be used for the product substrate are also very expensive but are becoming available
Stepping exposure is performed using a photomask manufactured by this method as a reticle.
There is a possibility that three-dimensional lithography on a large area substrate can be realized.
.
However, this indirect method has the following two problems shown in a and b.

a The exposure time becomes extremely long.
Binary pattern that usually draws wiring patterns on TFT-LCD substrates
The thickness of the resist used in the process is 0.5-1 μm, even 3 μm
Although it does not exceed, the pattern that is naturally formed in 3D lithography
The resist thickness must be greater than the depth of the resist.
Depending on the pattern to be formed, the rear projection display
In the case of a microlens array for clean use, a minimum of 30 μm,
Considering the maximum, about 100 μm is necessary.
The exposure time in binary patterning is about 5-30 seconds, so it is simple
Even if it takes only the film thickness magnification, it takes 500 to 3000 seconds for one exposure.
However, the light in the deep part by Lambert-Beer's law
The amount is attenuated by the amount absorbed before reaching the site, and the resist thickness
The required exposure time increases exponentially as the number increases.
When stepping exposure is performed in this situation, the first section
From the start of exposure until the last section exposure is complete
Will take a long time. For example, one exposure can be completed in one hour
Even so, using a 100mm reticle, the diagonal 70 inches wide
It takes 144 hours or more to expose the screen.
It must be exposed. The resist may change during this time
There is sex.

b Stepper has a problem of depth of focus.
In a mask exposure device other than a stepper,
The photomask is closely attached to the number or referred to as proximity
Parallel light is irradiated from the photomask side, with a micrometer gap in between.
The shadow of the mask pattern is projected onto the resist layer on the substrate surface. Therefore
There is no factor of resolution reduction other than the unsharpening of edges due to diffraction.
In the case of a stepper, the position of each pattern formed by multiple exposures
Substrate stage to reduce moving parts to the minimum necessary
Moves only in the XY direction. Therefore, the stepper
Contact exposure and proximity exposure used in other mask exposure systems
Not used. Because the movement of the substrate stage interferes with the position of the photomask
It is.
In steppers, optical masks are not placed near the substrate.
A system is used in which a mask pattern is imaged on the substrate surface by a system.

As a result, the substrate on which the mask pattern is drawn (called a reticle)
) Can be placed on the opposite side of the substrate with the optical system in between.
There is no interference.
However, this causes another problem. The resolution of the image pattern is
There is an effective distance in the Z-axis direction that does not decrease beyond the minimum line width.
Called degrees. The maximum depth of focus for large steppers for LCD substrates
3D lithography with a shape deeper than this
-Is impossible.

5). Laser ablation
When the surface of the material is irradiated with a narrowly narrowed laser beam, the area is very
Given a high density energy, it is rapidly evaporated and removed. This technology
Widely used for cutting and marking. The energy of this laser beam
3D pattern whose depth is modulated by scanning the substrate surface while modulating
Can be produced. Substrate size up to about 2m ²
It can be easily handled.
However, the ablated substrate surface is generally extremely rough and rough.
A flat surface cannot be obtained. The period of unevenness is constant
Although it is quite finely controlled, it is only slightly larger than the wavelength of visible light.
It will not get smaller. If such irregularities exist on the surface of the microlens array
In this case, scattering of image light occurs, and normal diffusion characteristics cannot be obtained.

  As described above, at present, even if a method for producing a highly controlled fly-eye lens exists, it cannot be produced on a large area substrate such as a screen for a rear projection display. Even though there is a means for performing microfabrication, it is currently impossible to precisely control the shape.

  In view of the present situation as described above, the present invention solves the roughening of the substrate surface, which is a drawback in laser ablation, and has a curvature controlled with extremely high precision on a large-area substrate, which has been impossible in the past. Another object of the present invention is to provide a method for manufacturing a microlens array matrix that can be fabricated from a unit cell.

  The gist of the present invention will be described with reference to the accompanying drawings.

  A method of manufacturing a master for forming an optical member provided with a large number of microlens arrays, wherein a photomask 2 having a mask pattern formed thereon is disposed above a substrate 1 and the photomask 2 is interposed therebetween. When a desired three-dimensional pattern corresponding to the mask pattern is formed on the substrate 1 by irradiating the laser beam 3, the spot size of the laser beam 3 is enlarged and transmitted through the photomask 2, and then the laser beam 3 is directed to a method for manufacturing a matrix for a microlens array, wherein a part of the surface of the substrate is evaporated by reducing the size of the spot 3 and irradiating the surface of the substrate 1.

  The method for manufacturing a microlens array matrix according to claim 1, wherein a binary mask is employed as the photomask 2.

  3. The microlens array matrix according to claim 2, wherein a plurality of binary masks are used, and the plurality of binary masks have different mask patterns. This relates to the manufacturing method.

  Further, in the method for manufacturing a microlens array matrix according to claim 1, a gradation mask having a gradation of transmittance at the light shielding portion is employed as the photomask 2, and the gradation mask is made of metal, The present invention relates to a method for manufacturing a matrix for a microlens array, wherein the transmittance gradation is formed by changing the thickness.

Further, in the method for manufacturing a microlens array matrix according to claim 4, the following steps (1), (2), (3), and (4) are sequentially performed on the gradation mask. The present invention relates to a method for manufacturing a matrix for a microlens array, which is manufactured by the method described above.
Step (1): A step of applying an electron beam resist to the surface of the metal film provided on the substrate.
Step (2): The electron beam resist surface is scanned by modulating the electron beam dose.
Irradiating the electron beam;
Step (3): develop and remove a portion irradiated with a dose of electron beam exceeding the threshold value to obtain a thickness gradation
Forming an electron beam resist layer.
Step (4): Electron beam resist with gradation added by reactive ion etching
A process for removing the thickness layer and transferring the thickness gradation of the electron beam resist layer to a metal film.
About.

  2. The method for manufacturing a matrix for a microlens array according to claim 1, wherein a gradation mask having a gradation of transmittance at the light shielding portion is employed as the photomask 2, and the gradation mask absorbs an electron beam. The electron beam is applied to a glass having a layer in which the transmittance of light including a specific wavelength changes, and there is a range in which the absorbed electron dose and the transmittance of light including the specific wavelength correspond to one to one. The present invention relates to a method of manufacturing a mother die for a microlens array, which is manufactured by modulating and irradiating to give gradation to exposure light transmittance.

The method for manufacturing a microlens array matrix according to any one of claims 3 to 6, wherein a plane perpendicular to the main plane of the microlens array matrix and the microlens array matrix are intersected. A figure formed by intersecting a plane parallel to the main plane passing through a point obtained by dividing the longest intersection of the lines into a length represented by the following formula (1) and the matrix surface for the microlens array Is used as a gradation pattern, and the present invention relates to a method for manufacturing a matrix for a microlens array.
Cs / (g−1) × (δ + 1) Formula (1)
Where Cs: a plane perpendicular to the main plane of the matrix for the microlens array, and the microlens
The length of the longest line of intersection on the matrix surface for the array
g: Number of gradations
δ: A number from −0.2 to 0.2.

  Since the present invention is configured as described above, there is provided a microlens array matrix that enables fabrication of a microlens array matrix composed of unit cells having a curvature controlled with extremely high precision on a large-area substrate. It becomes a manufacturing method.

  An embodiment of the present invention considered to be suitable will be briefly described with reference to the drawings, showing its effects.

  A photomask 2 on which a mask pattern is formed is disposed above the substrate 1, and a laser beam 3 is irradiated onto the substrate 1 through the photomask 2, and the laser beam 3 reflecting the mask pattern of the photomask 2. Thus, the surface of the substrate 1 is processed into an arbitrary shape.

  At this time, the energy density of the laser beam 3 is controlled by expanding and contracting the beam spot of the laser beam 3 to reflect the mask pattern of the photomask 2 without applying excessive energy to the photomask 2 and the substrate 1. The surface of the substrate 1 can be processed by the laser beam 3 thus made, and thus a matrix for a microlens array comprising unit cells having a curvature controlled with extremely high precision on a large area substrate, which has been impossible in the past. This is a method for manufacturing a matrix for a microlens array that enables fabrication of a microlens array.

  Specific embodiments of the present invention will be described with reference to the drawings.

  This embodiment is a method for producing a smooth curved surface by laser abrasion by irradiating the surface of the substrate 1 with a laser beam 3 that has passed through a photomask 2.

  Processing of the surface of the substrate 1 with the laser beam 3 is performed by a commercially available laser processing machine. Specifically, this laser processing machine includes a laser oscillator 4 that generates a laser beam 3, an optical system that expands the diameter of the laser beam 3, a photomask 2 on which an arbitrary mask pattern is formed, and the photomask 2 A mask holder provided, a mask positioning mechanism for positioning the photomask 2, and an optical system 6 for reducing the diameter of the laser beam 3 that has passed through the photomask 2 and projecting it onto the surface of the substrate 1 to be processed are provided. Is.

  The light source laser is appropriately selected according to the required resolution, the material of the substrate to be processed, and the material of the photomask 2. When higher resolution is required, it is preferable to use a light source capable of obtaining output light with a short wavelength. Accordingly, Nd: YAG harmonics and excimer lasers are suitable for the laser light source applied to this embodiment.

  Laser ablation is a phenomenon in which a chemical bond in the material constituting the substrate is cut by absorbing light on the surface of the substrate irradiated with laser light of a specific wavelength, and evaporated as a segment with a small molecular weight. . Therefore, in order to perform processing, it is necessary to use a laser light source having a wavelength that is absorbed by the substrate material. In general, most materials used for such microfabrication, such as glass, ceramics, or plastics, have absorption in the ultraviolet region, so an ultraviolet laser is used for any of those materials. It will be.

  The material of the photomask 2 also needs to consider the relationship with the wavelength. That is, the photomask 2 is a part for selectively shielding the light used for processing and allowing the light to selectively reach the surface of the substrate 1, and for this purpose, a portion that completely absorbs the light from the laser light source. The transparent part is arranged in one mask.

  Although the beam spot size of the light source is not particularly limited, since the total processing time is expressed by the following formula (2), the processing time can be shortened when the number of exposure areas, that is, the number of divisions is small. Therefore, it is better to increase the beam spot size on the surface of the substrate to be processed.

t ₀ × g × d + α Formula (2)
However, t ₀ is a single exposure time, g is the number of gradations, d is the number of exposure areas, and α is a time required for operations other than the exposure process (mainly movement between exposure areas).

Desired input energy density at the surface of the substrate to be processed 0.1 J / cm ² or more, 10J / cm ² or less is appropriate, the rate of abrasion when the input energy density is lower than 0.1 J / cm ² Is too small and takes a long time, and if it is higher than 10 J / cm ² , the surface becomes so rough that it can be used for optical applications as in the case of abrasion by beam scanning that is usually performed. do not become.

  In addition, it is preferable to set the substrate material to a relatively low energy density when the material is easy to evaporate such as plastic, and to a relatively high energy density when the substrate material is difficult to evaporate such as glass or ceramic material. .

  The most common material used for the photomask substrate is glass. Therefore, when the photomask is directly irradiated with the laser beam 3 applied to the substrate surface, the photomask itself is ablated and acts as a mask. Will not be able to.

  Accordingly, in order to reduce the energy density, the laser beam 3 emitted from the light source is once enlarged in cross-sectional area by the optical system 5 and passed through the photomask 2, and then again reduced in cross-sectional area by another optical system 6. It is necessary to irradiate the surface of the substrate 1. The magnification of the beam cross-sectional area varies depending on the energy density on the surface of the substrate 1, but is generally preferably between 2 and 100 times the cross-sectional area on the substrate surface. If it is less than 2 times, the energy density applied to the photomask 2 is too high and the temperature rises too much even if the photomask 2 does not cause abrasion, and if it is more than 100 times, the photomask 2 is necessary. Inconveniences such as a large area and insufficient number of gradations occur. In addition, the beam spot size before expansion emitted from the light source and the beam spot size when irradiating the substrate surface need not be the same.

  The exposure unit surface, that is, the beam cross-section or photomask effective area must be exposed without excessive or deficient, so that the entire surface can be filled without overlapping. is required. Such shapes vary from relatively simple figures such as squares, rectangles, rhombuses, parallelograms, hexagons having two or more symmetry axes to complex shapes shown in FIGS. 2 (A) and 2 (B). Can be used.

  The mask pattern drawn on the photomask 2 reflects a cross section obtained by cutting the arrangement of the curved surface pattern of the matrix for microlens array to be produced at a predetermined height by a plane parallel to the main plane. It does not match the shape. The light shielding area is narrower or wider than that determined by the corresponding cross section. The shape of the mask pattern is determined so that the shape of the portion of the surface of the substrate 1 scraped by the laser beam 3 formed by the mask pattern is equivalent to the curved surface pattern of the matrix for microlens array to be produced.

  The height of the array of curved surface patterns of the microlens array master block to be produced by a plane parallel to the main plane depends on the number of steps (the number of gradations) of the stepped pattern approximating the curved surface pattern.

  In consideration of the relationship between the shape of the edge of the staircase pattern formed during abrasion and the size of the step, the number of gradations is preferably reduced as long as no inconvenience occurs under the conditions used. Since the number of gradations is equal to the number of masks, an increase in the number of gradations means that a large number of masks must be produced, and further, the number of times of irradiation of one exposure region also increases. That is, an increase in the number of gradations results in an increase in mask manufacturing cost and an increase in process time.

  That is, on the substrate surface irradiated with the laser beam 3, the edge portion of the beam becomes a boundary between the portion that is ablated and evaporates and the portion that does not evaporate, and the boundary portion is the wavelength, input power, pulse width, substrate at that time It has a specific shape that depends on the type of material.

  For example, when the temperature rise is not so large, a steep step shape is shown, but when the temperature rises, the corner is melted and rounded, and when it is severe, it becomes a gentle slope. Such a smoothing phenomenon seen when the temperature rise is large brings about an advantage that a smooth curved surface can be produced with a small number of gradations, that is, the number of masks.

  However, when such a smoothing effect is too great, there arises a disadvantage that a deviation from the design of the produced curved surface shape becomes large. Therefore, it is necessary to determine the number of masks and the laser beam irradiation conditions in consideration of the balance between the advantages of reducing the number of masks due to the smoothing effect and the disadvantages of deviation from the design.

  In addition, when it is necessary to keep the deviation from the design of the created curved surface shape within a very small range, since the smoothing effect due to temperature rise cannot be expected, the completed shape was composed of stepped side surfaces instead of smooth curved surfaces Become a thing. In such a structure, if the effect of scattering or diffraction caused by the step structure becomes a serious problem, a smoothing process is performed by melting the step portion by heat treatment later or filling the step portion with a coating. Such inconvenience can be reduced by applying.

  The number of gradations is determined in consideration of the above, and the light shielding pattern in each gradation is determined in consideration of the shape of each cross section and the nature of the abrasion process used at that time, mainly the smoothing effect from the temperature rise.

  Using the photomask 2 having such a gradation, abrasion is caused while adjusting the exposure amount on the substrate surface in the following procedure. There are two methods, one is to achieve all gradations with a single exposure, and the other is to draw each gradation on a separate mask and use each mask at least once to superimpose at least the number of gradations. May be.

  In order to realize all gradations by one exposure, a photomask 2 in which each gradation is drawn on one sheet is used. Such a photomask 2 can be manufactured by a method of controlling the thickness of the metal film or the amount of coloring of special photochromic glass.

  The thickness control of the metal film is performed as follows. First, a material that transmits exposure light such as glass is used as a substrate, and a metal film such as Cr is formed thereon by sputtering or the like. An electron beam resist is formed (coated) thereon and scanned while modulating the electron beam dose to irradiate the entire surface of the electron beam resist. The modulation pattern corresponds to the thickness pattern of the metal film. By developing this and removing a portion irradiated with a predetermined amount or more of the electron beam, a stepwise resist pattern (thickness gradation) corresponding to the target curved surface pattern of the microlens array matrix is obtained. By etching the electron beam resist and the metal film together by RIE (Reactive Ion Etching) from the resist surface in a direction perpendicular to the matrix for the microlens array, the thickness gradation of the electron beam resist layer can be changed to Cr or the like. Transfer to the thickness gradation of the metal film.

  That is, a substrate with a Cr film provided with an electron beam resist film having a developed relief on the surface is placed in a vacuum chamber, and the substrate is formed by two flat plate electrodes parallel to the main plane of the mother mold for the microlens array to be manufactured. After depressurizing the exhaust gas between the electrodes, RF (Radio Frequency; high frequency in the radio wave region; 10 to 100 MHz) is applied between these electrodes with a DC voltage or a DC bias while introducing an inert gas and a reactive gas.

  An inert gas is introduced to stabilize the plasma generated in the vacuum chamber. Basically, rare elements such as Ar and Xe are used because it is better not to react with the inner wall of the vacuum chamber or other members. As the reactive gas, a gas having appropriate reactivity with the electron beam resist and the metal is used. Since electron beam resists are usually composed mainly of organic polymers, they are highly reactive to electron beam resists, that is, when increasing the etching rate of electron beam resists, oxygen-containing gas is used, and in many cases oxygen itself is used. Introduce. Conversely, when increasing the etching rate of the metal film, a gas containing halogen is introduced. Such gases include tetrafluoromethane, trifluoromethane, tetrachloromethane, trichloromethane, and the like. These gases can be used alone or in combination.

  Reactive and non-reactive etching by ions generated by ionizing the gas from the electron beam resist surface having a fine curved surface by selecting appropriate gas composition and pressure, discharge voltage, bias voltage, etc. Thus, the thickness decreases while maintaining the surface shape, the surface of the metal film is exposed from the portion having the smallest electron beam resist thickness, and then the metal film is etched. When a gas composition having a higher metal etching rate is used, the unevenness on the surface of the metal film becomes steeper than the unevenness on the surface of the electron beam resist, and vice versa.

  An antireflection film can be provided on the surface of the gradation mask manufactured as described above. The antireflection film on the mask surface has a function of avoiding adverse effects due to stray light, but is not necessarily required if an optical design is made so that the reflected light does not return.

  The photomask 2 using a special photochromic glass in which the coloring amount is controlled is manufactured as follows. This glass is called HEBS (High Energy Beam Sensitive) glass, and is a glass that causes coloration in a wide band by an electron beam, and can be obtained from Canon Materials, Inc. More are sold. It can be produced by scanning the glass while scanning the electron beam while modulating the electron beam in the same manner as in the case of producing the metal gradation mask.

  Each gradation is drawn on a separate mask, and the gradation mask used in the method of superimposing at least the number of gradations by using each mask at least once is generally a binary mask that expresses each gradation. It can be manufactured by a typical photomask manufacturing method.

  Needless to say, the photomask 2 used in this case is a binary mask used in general thin film patterning, and the positional relationship between a plurality of binary masks corresponding to each gradation is as accurate as possible. Need to control. The problem of positioning accuracy is always a more important factor in this embodiment, including the problem of positioning accuracy between exposure areas, which will be described later. The required level is higher than the joint positioning accuracy. This is because the alignment error between exposure areas can be corrected in a later process, while the positioning error between gradations within a single exposure area directly affects the shape accuracy of the curved surface to be produced. . An allowable amount of positioning error between gradations is determined by the size of the microlens array unit cell to be manufactured, and is preferably within 1% of the unit cell size, specifically within 0.1%.

  Further, it is easier to produce a deeper shape by ablating once using the photomask 2 for each gradation, compared to a method of ablating all gradations once. In the case of the photomask 2 in which all gradations are drawn on one mask blank, the limit of the depth of the curved surface pattern that can be produced is determined by the dynamic range of the transmittance and the output of the laser transmitter used. In the case of performing abrasion for each gradation, there is no limit to the depth that can be produced. Of course, if the pattern becomes deeper, the number of times of irradiation with the laser beam 3 increases, so that the time required for production becomes longer.

  As an important part of the present embodiment, there is a method of determining brightness at each gradation.

  Generally, when gradation display is performed within a certain range, the darkest level and the brightest level are divided at equal intervals in brightness, and each level is set as the level of each gradation. This relationship is generally expressed by the following formula (3). In applications that are seen with the naked eye, i.e., applications that place importance on visual fidelity such as printed matter, gamma correction, i.e., correction of visual characteristics, and in many cases, bias in the shape of an S-curve.

_{L 0 + (L h -L 0} ) / (gl-1) (3)
However, L ₀ is the lowest level of luminance, L _h is the highest level of luminance, and gl is the number of gradations.

  However, when gradation division is performed by this method when creating a curved surface by abrasion, the resolution at a portion where the angle between the tangent line and the main plane is small as shown in FIG. 3 decreases. On the contrary, in a region where this angle is large, processing is performed with a resolution higher than the apparent resolution depending on circumstances, which is practically impossible.

  Therefore, in this embodiment, as shown in FIG. 4, the gradation division of the gradation mask to be used is the intersection of the main plane of the microlens array master block to be produced and the microlens array master block. It was decided that the intersection line having the longest length was divided into approximately the number of gradations minus one.

  However, this dividing point does not necessarily have to be an equal dividing point, and may be almost in the vicinity of the dividing point. Therefore, the length is within the range of about 20% from the equivalence point.

  As described above, by determining the gradation dividing point based on the shape of the curved surface, the resolution can be made uniform at all gradations, so that the fidelity to the design of the produced curved surface pattern is improved.

  In this embodiment, any of the three types of photomasks 2 as described above can be used. When a photomask 2 in which all gradations are drawn on one photomask 2 is used, the surface of the substrate 1 is irradiated at least once through the photomask 2 and then the next exposure region. The same exposure process is performed on the above. When separate photomasks 2 corresponding to the respective gradations are used, at least the number of masks, that is, the steps of exposing the photomask 2 to the same exposure region on the surface of the substrate 1 while maintaining a specific positional relationship. After performing the logarithm times, the process proceeds to the next exposure area.

  As described above, if necessary, the exposure area is filled and exposed without gaps over the entire surface of the substrate to produce a target microlens array matrix. At this time, the allowable amount of positioning error at the joint between adjacent exposure regions is preferably １ ／ or less, preferably 1/50 or less of the unit cell size, depending on the size of the unit cell to be produced. The reason why the tolerance for positioning error at the joint between adjacent exposure areas is larger than the tolerance for positioning error between gradations in a single exposure area is as follows.

  The positioning error between gradations in a single exposure area affects the curved surface shape that is immediately produced. Therefore, even if the error in this case is large, it is necessary to suppress it to 1% or less, preferably 0.1% or less of the unit cell size. On the other hand, the positioning error at the joint between adjacent exposure regions does not affect the curved surface shape to be produced. Of course, when the joint between adjacent exposure areas is inside the unit cell, a joint will be formed in the curved surface. Therefore, when the exposure area is designed in this way, the positioning error at the joint between the adjacent exposure areas is determined. In addition, an extremely small numerical value is required as in the case of the positioning error between gradations in a single exposure region. However, the exposure area can be designed so that the junction between adjacent exposure areas does not cross. When such a design is performed, a considerably wide seam portion is formed by forming a light shielding film described later. Even if it exists, the influence can be removed.

  However, the joint portion is a light shielding portion, and if the ratio of the area of this portion to the entire microlens array increases, the proportion of the incident light absorbed by the sheet also increases. That is, the light utilization efficiency is reduced. Therefore, it is better that the width of the seam portion is not too wide. From the above, it is preferable that the allowable amount of positioning error at the joint between adjacent exposure regions is 1/5 or less, preferably 1/50 or less of the unit cell size.

  By passing through the above processes, a smooth three-dimensional pattern formed by laser abrasion is completed on the substrate surface. Using this as a mother die, a mother die for molding a microlens array is manufactured.

  Metals and synthetic resins are generally used as mold materials. When using a metal, a method generally called electroforming is used. A thick metal film is grown by electrolytic plating on a conductive thin film.

  For the formation of the conductive thin film, silver mirror reaction, vacuum deposition, sputtering, or the like is used.

  Ni is generally used for the metal deposited by electrolytic plating. In the case of a synthetic resin, shape transfer is performed using a room temperature curable silicone resin generally called SMC (Silicon Molding Compound).

  In this way, the fine curved surface shape of the substrate surface is transferred to the mold, and the microlens array sheet is molded. A non-solvent curable resin composition is poured onto the mold, and a microlens array substrate is placed thereon and cured. The thickness of the microlens array substrate is set to about the focal length of the microlens.

  After the microlens array is cured, the mold is released, and a film made of a photosensitive adhesive is laminated on the surface of the microlens array substrate where the microlens array is not formed. Only the portion that interferes with the optical path of the photosensitive adhesive is exposed and cured by the light collecting action of the lens array.

  As the photosensitive adhesive, those that are generally available such as a type that crosslinks by a photoradical reaction, a type that crosslinks by an acid-base reaction by a photocation, and the like can be used.

  Thereafter, a black powder or a fragile black film is laminated and adhered to a portion that does not interfere with the remaining optical path of the adhesive, thereby forming a light shielding layer (black matrix). As a result, the surface of the microlens array sheet that is transparent only on the optical path and does not become an optical path can be made black, so that the contrast of the screen can be increased.

  Further, a back sheet is laminated and pasted thereon to protect the black matrix. If an antireflection film is formed on the surface of the microlens array or the back sheet, stray light due to reflected light does not occur and reflection loss can be reduced.

  Since the present embodiment is as described above, it is possible to avoid the contradictory relationship between the resolution and the processing speed, which has been unavoidable in the conventional laser ablation technology. That is, the resolution is determined by the resolution of the photomask by enlarging the laser beam and irradiating the substrate surface through the photomask. Further, since the area of the region to be ablated at a time is increased, ablation can be caused without scanning at high speed. Accordingly, since the amount of energy input per unit time × area becomes small, the surface material that has been ablated does not rapidly evaporate, and the ablated surface becomes smooth. Further, in the present invention, the thickness of each layer is not equalized in the gradation division method performed in the gradation exposure using the photomask, but the length of the curve intersecting the plane perpendicular to the main plane and the curved surface is By performing substantially equal division, it is possible to realize gradation division with substantially uniform resolution over the entire curved surface. From the above, according to the present embodiment, it is possible to manufacture a large-area microlens array having a smooth curved surface with a highly controlled shape.

It is a schematic explanatory drawing of a present Example. It is an example of the shape of the exposure area | region of a present Example. This is an example in which gradation division is performed using contour lines. This is an example of gradation division based on a curved surface shape.

Explanation of symbols

1 Substrate 2 Photomask 3 Laser beam

Claims (7)

  A method of manufacturing a master for forming an optical member provided with a large number of microlens arrays, wherein a photomask having a mask pattern formed thereon is disposed above a substrate, and a laser beam is emitted through the photomask. When forming a desired three-dimensional pattern according to the mask pattern on the substrate by irradiating, the laser beam spot size is enlarged and transmitted through the photomask, and then the laser beam spot size is reduced. A method for manufacturing a matrix for a microlens array, wherein a part of the substrate surface is evaporated by irradiating the substrate surface.   2. The method for manufacturing a microlens array matrix according to claim 1, wherein a binary mask is employed as the photomask.   3. The method of manufacturing a microlens array matrix according to claim 2, wherein a plurality of binary masks are used, and the plurality of binary masks have different mask patterns. Method.   2. The method of manufacturing a matrix for a microlens array according to claim 1, wherein a gradation mask having a gradation of transmittance at a light shielding portion is adopted as a photomask, and the gradation mask is made of metal and the thickness thereof is changed. A method for manufacturing a matrix for a microlens array, wherein the gray scale of the transmittance is formed. 5. The method of manufacturing a microlens array matrix according to claim 4, wherein the gradation mask is manufactured by sequentially performing the following step (1), step (2), step (3), and step (4). A method for manufacturing a mother die for a microlens array, which is characterized by the above.
Step (1): A step of applying an electron beam resist to the surface of the metal film provided on the substrate.
Step (2): The electron beam resist surface is scanned by modulating the electron beam dose.
Irradiating the electron beam;
Step (3): develop and remove a portion irradiated with a dose of electron beam exceeding the threshold value to obtain a thickness gradation
Forming an electron beam resist layer.
Step (4): Electron beam resist with gradation added by reactive ion etching
A process for removing the thickness layer and transferring the thickness gradation of the electron beam resist layer to a metal film.
About.   2. The method for manufacturing a matrix for a microlens array according to claim 1, wherein a gradation mask having a gradation of transmittance at a light shielding portion is adopted as a photomask, and the gradation mask is specified by absorbing an electron beam. A glass having a layer in which the transmittance of light including a wavelength is changed, and having a range in which the absorbed electron dose and the transmittance of light including the specific wavelength have a one-to-one correspondence, A method for manufacturing a matrix for a microlens array, which is produced by irradiating and providing gradation of exposure light transmittance. The method for manufacturing a matrix for a microlens array according to any one of claims 3 to 6, wherein a plane perpendicular to the main plane of the matrix for the microlens array and an intersection of the matrix for the microlens array A figure formed by intersecting the plane parallel to the principal plane passing through the point obtained by dividing the longest intersection line into the length represented by the following formula (1) and the matrix surface for the microlens array is a floor. A method of manufacturing a mother die for a microlens array, characterized by being used as a tone pattern.
Cs / (g−1) × (δ + 1) Formula (1)
Where Cs: a plane perpendicular to the main plane of the matrix for the microlens array, and the microlens
The length of the longest line of intersection on the matrix surface for the array
g: Number of gradations
δ: A number from −0.2 to 0.2.

Images